With the Institute of Medicine's call for a more evidence-based, efficient, patient-centered healthcare system, this study proposes to test the impact of a new mechanism for eliciting informed consent from patients undergoing percutaneous coronary intervention (PCI) on patients'comprehension of procedural risks/benefits and upon clinicians'use of effective strategies to minimize the risk of bleeding at the time of PCI. Bleeding is the most common, non-cardiac complication of PCI and is strongly associated with myocardial infarction (heart attack), stroke, increased length of stay/costs and death. While important adjunctive therapies exist to minimize the risk of bleeding, in current practice, these therapies are most often applied to patients at low risk for bleeding and least often applied to those at the highest risk - a risk-treatment paradox. Our proposal seeks to clarify patients'risk of bleeding at the time of PCI so that these adjunctive therapies can be most rationally applied to those at highest risk to support safer, more cost-effective, patient-centered care. To accomplish this, the infrastructure of the informed consent process at participating study centers will be transformed by deploying a novel, web-based system - the Personalized Risk Information Services Manager (PRISM) - that generates individualized consent forms with estimates of risks and outcomes using validated multivariable models from contemporary practice. Using a pre/post study design at 6 institutions, supplemented with a case control matching methodology, the impact of this new mechanism for eliciting informed consent from patients undergoing PCI will be evaluated by explicitly testing whether PRISM: a) improves the quality of the informed consent process, b) supports the more rational use of Bleeding Avoidance Therapies among those at the highest risk for bleeding, c) decreases bleeding events at the time, or within 1 month, of PCI and d) supports a more cost-effective model of PCI. This project tests a paradigmatic shift in the practice of medicine that is directly aligned with the strategic directions of the NHLBI to translate new knowledge into routine clinical care. Given the unprecedented investment of NHLBI and the ARRA Stimulus Package in supporting comparative effectiveness reIT infrastructure, whether it can improve patients[unreadable] experience with health care, lead to more rational medical decision-making or whether it is even economically feasible to redesign healthcare delivery in this direction. Towards this end, our team has developed the Personalized Risk Information Services Manager (PRISM;www.cardioscience.org;username: guest, password: guest), an entirely new paradigm for delivering patient-specific outcomes projections at the interface of patient care.3,4 A particularly compelling opportunity to test the impact of evidence-based, patient-centered care is at the time of acquiring informed consent for percutaneous coronary interventions (PCI). While >1,000,000 PCI procedures are performed annually, and although multiple riskprediction models are available,5-10 these models are never used to present patients with estimates of their risk for PCI complications or to guide periprocedural management. A particularly important opportunity to influence care at the time of PCI is management of bleeding risk. Bleeding is common, morbid, and costly, occurring in 2-6% of patients undergoing PCR with wide institutional variability. Major bleeding events also result in a 3-4 day increase in length of stay,11 in turn increasing hospitalization costs by $6,000-8,000.12 Bleeding is also associated with morbid events including non-fatal MI, stroke, and death. Importantly, peri-procedural bleeding risk is modifiable, as there are established bleeding avoidance therapies (BATs) including effective anticoagulation treatments (e.g., bivalirudin), vascular closure devices, and radial access techniques that can mitigate the risk. However, these interventions can be costly, mandating a rational strategy to apply them in higher risk patients, while avoiding them in patients at lower risk. In partnership with the National Cardiovascular Data Registry(R) (NCDR(R)), our group developed and validated a clinical risk model to estimate individual patient risk for post-PCI bleeding using a large nationally representative database of >300,000 patients from >400 hospitals.13 To address this opportunity to improve the quality of informed consent, and to realize the Institute of Medicine[unreadable]s goal of delivering safer, more efficient, evidence-based, patient-centered care, we will 1) implement PRISM-generated, individualized informed consent documents at 6 high-volume PCI hospitals and 2) test its association with the quality of informed consent. We will also examine the association of PRISM-generated consents with 3) the use of BATs in those with the greatest potential to benefit (i.e. those at the highest risk for bleeding) Finally, we will integrate all of these goals by formally assessing the cost-effectiveness of PCI after introducing PRISM into practice. The process of performing PCI (solid lines) and the Specific Aims of this project (dashed lines) are diagramed in Figure 1 and outlined below: Specific Aim 1: Identify the barriers in implementing individualized informed consents in routine clinical care. If Specific Aims 2-5 demonstrate that the PRISMgenerated consent forms elevate the quality, efficiency and cost-effectiveness of PCI, there will be a need to disseminate this approach - ideally through the American College of Cardiology's NCDR, which is already deployed in >800 hospitals, To facilitate future dissemination, we will study the challenges in implementing PRISM at our 6 study sites, including nurse, physician and staff training, risk management and informatics barriers that we overcome. Throughout our implementation, we will develop tools and educational materials that will ease future efforts to integrate more educational, evidence-based, individualized consent forms in clinical practice. Specific Aim 2: Test the impact of PRISM on the quality of informed consent: It is widely recognized that current informed consent documents - written in a vague, complex, legalese style - fail to communicate the relevant information to patients in a fashion that they can understand.14-16 While we have designed PRISM to imbed the output of its prediction models in a more educational informed consent document, the ability of this document to meet the goals of informed consent has not been definitively demonstrated. We will formally test the [unreadable]quality[unreadable] of PRISM-generated consents to educate patients regarding the risks and benefits of PCI as compared with traditional consent forms. Specific Aim 3: Test the impact of PRISM on the use of Bleeding Avoidance Therapies (BATs): At the time of PCI, physicians must choose the best strategies to minimize the risks of bleeding, a major procedural complication. In a comparative effectiveness analysis of the same data upon which we built our bleeding risk model,13 we found that patients at the lowest risk for bleeding were less likely to receive no BATs (bivalirudin or closure devices) as compared with those at moderate and high risk (38% vs. 42% vs. 49%, p<0.001) and more likely to receive both closure devices and bivalirudin than higher risk groups (15% vs. 13% vs. 10%, p<0.001);a dramatic demonstration of the risk-treatment paradox underscoring the need to transparently display patients[unreadable] risks at the time of treatment to support more rational decision making. The impact of PRISM on the use of BATs, as a function of pre-procedural risk, will be tested within study sites using a pre-post study design, after adjusting for secular trends with the NCDR Cath/PCI registry (matching study sites 5:1 on hospital characteristics, including PCI volume and BAT use at the time of PRISM[unreadable]s introduction), so that we can define the association of PRISM with changes in peri-procedural bleeding management. Specific Aim 4: Test the impact of PRISM on bleeding outcomes after PCI: We will prospectively monitor patients throughout their hospitalization and by a 1-month telephone call to assess for peri-procedural bleeding. By increasing the use of BATs in high-risk patients, we hypothesize that we will significantly decrease the occurrence of bleeding across the study population.. Specific Aim 5: Create a formal economic model of the cost-effectiveness of DES after implementing PRISM-generated informed consents. Enhancing the infrastructure for obtaining better informed consent (Specific Aim 1) costs money. To offset these costs, it is imperative to estimate the economic benefits of PRISM. We also anticipate that our strategy will lead to a greater use of BATs in high-risk patients and a lower use in low-risk patients. In a cost model of >1800 patients conducted at our institution, the average cost of bivalirudin use in low risk patients was $724/patient, whereas its use in moderate- and high- risk patients was associated with savings of $498 and $1595, respectively. Our research team has extensive experience in the construction of economic models in PCI. Critical decision nodes for the costeffectiveness of BATs are the use of these therapies and the occurrence of observed bleeding events. By not only surveying in-hospital, but also conducting 1-month follow-up, we will have complete assessment of bleeding events and be able to efficiently estimate the impact of PRSIM on BAT use and cost. We will use the results from Specific Aims 3 and 4 to define the incremental improvements in these parameters associated with PRISM use. We will then estimate the cost-savings associated with more rational BAT use to test the hypothesis that PRISM will improve the cost-effectiveness of PCI.